Three Steps to Spring Clean Your best BI Platform

Stumbling through an information portal cluttered with content choices and complex navigation can be a confusing experience. It may seem like the answer you want is out-of-reach. This confusion is often a daily experience for many business intelligence platform users. They can become exasperated by the difficulty of finding and reporting on the information their jobs require. This difficulty-of-use is a primary contributor to poor BI adoption. 

However, there are actions we can take as BI practitioners to improve the user experience—and adoption. Moreover, the outcomes of our actions can be measured in data showing how our platform is used.

In this blog post, you’ll discover techniques to get a clearer picture of your users’ experience and take action to improve it.

Step 1: “BI-on-BI” - Enabling Your BI platform to Report on Itself

Captured inside every BI platform is a variety of user behavioral pattern data, system logs, content metadata, and other information that often goes unseen and unused. If you tap into this information, you can unleash potentially valuable insights and leverage them to sharpen the user experience, and ultimately increase adoption of the BI platform.

Understand and fully enable your BI platform’s auditing and logging capabilities

Enable your platform’s auditing features to the maximum extent because this information helps you derive insights about how users are interacting with content.  (Of course, be mindful of your storage limits when you do this.) 

By default, most BI vendors log at minimal levels to conserve storage space. As a result, many large BI implementations at Fortune 1000 companies have thousands of reports and users but little insight into who is logging into the system. The administrator don’t know which content is being accessed, simply because they hadn’t fully enabled auditing. And they hadn’t bothered to build reports or dashboards to answer these questions.

At a minimum, your platform should collect and maintain at least a year’s worth of data. This data should include:

  • user logon/logoff events
  • session duration
  • reports and dashboards that were accessed and by whom
  • queries sent to the database and the resulting query/report performance times
  • any interactions or clicks users perform, such as changing filter criteria on-the-fly
  • error messages, warnings, or resource contention issues

Depending on your platform, even more information may be available.

Additionally, if your vendor doesn’t provide a relational auditing database schema, you should set up processes that collect logging events into a database so you can analyze this information. Consult your vendor’s user community and support teams for their advice about the best tools and methods to maximize the inward-looking capabilities of your BI platform. 

Pull the weeds! 

Look for the weeds in your data such as reports that are outdated or no longer used. Archive or delete the weeds. Just as weeding a garden helps you see the healthy plants, removing this old data can clear a user’s visual navigation experience.Weed-pulling can also accelerate system maintenance such as upgrades or backups. Use the auditing repository to determine which reports can be safely archived or deleted. If a report hasn’t been accessed in over a year, chances are good that it can be safely archived. Create auditing reports to help you identify your “bottom n least-used artifacts” and use that information to till the garden.

Step 2: Use Data to Help Optimize the User Experience

After your BI system is relatively healthy with production-quality, active content, you’ll want to focus on optimizing the paths users take to find those objects, and how easily they can interact with them.
Rank reports and push popular content to the top

A “top n most-used reports” auditing trail will help you identify artifacts that are good candidates for priority placement. Some BI platforms allow objects to be accessed in two ways: by the traditional folder hierarchy and also by keywords.

But not every implementation leverages this feature. 

If your platform supports content keywords, consider setting them up to provide a simpler way for users to find reports related to a particular subject, purpose, and/or data source.

Analyze the custom parameters users choose and rank them near the top of menus

A classic example of this is a Country parameter in a report, where your auditing data might reveal that 80% of the time, users enter “United States” in the parameter value. Instead of making the user scroll to the bottom of a list of 50 countries they seldom select, consider ranking United States at the top of the list to smooth their search experience. 

Many times, by evaluating which parameter values users select most frequently, you’ll find that 5-20% of the values are used 80% of the time. Where possible, create insights around this and enhance your report designs to minimize clicking or scrolling past seldom-used parameter values. 

Even better (and if your platform allows it), create parity between your data security model and the parameter values users see in the first place. If users in your Germany division aren’t allowed to see data in North America, why bother cluttering their filter choices with values that will never return them data?

Create context-aware links within reports to quickly connect users with related reports

Your auditing data (and practitioner's intuition) may reveal that certain reports are commonly consumed together. Consider implementing context-aware links within reports.  These links can launch other related reports that contain pre-applied filter criteria to smooth the user experience. This technique is similar to the recommendations widgets on leading consumer shopping and media platforms.

Hopefully these BI-on-BI techniques help you improve the experience for your users.  The same BI-on-BI data you use to help users should, with time, reveal the results of your hard work. You’ll be able to quantify the user experience improvements in terms of reducing the number of clicks users need to arrive at report, increasing the frequency with which they log on to the system and much more! 

Have you found other helpful techniques for cleaning up and optimizing your BI platform?  Let me know in the comments.